基于单目和双目视觉信息的全参考立体图像质量评价模型  

Full-Reference Model for Stereoscopic Image Quality Assessment Based on Monocular and Binocular Visual Information

作　　者：王宽 杨环 潘振宽 司建伟 WANG Kuan;YANG Huan;PAN Zhenkuan;SI Jianwei(School of Computer Science and Technology,Qingdao University,Qingdao,Shandong 266071,China)

机构地区：[1]青岛大学计算机科学技术学院,山东青岛266071

出　　处：《计算机工程》2022年第2期207-214,223,共9页Computer Engineering

基　　金：中国博士后科学基金(2017M622136);山东省重点研发计划(2019GGX101021);青岛市应用研究项目(2016025)。

摘　　要：在立体图像质量评价领域,有效地模拟人类视觉系统对图像质量进行评价具有重要意义,考虑到人眼的视觉感知特性,基于单目和双目视觉信息构建一种立体图像质量评价模型MB-FR-SIQA。采用基于结构相似性的立体视差算法得到参考和失真立体图像的视差矩阵,结合Gabor能量响应图、显著性图和视差矩阵生成中间视图,并优化左右眼加权系数计算方法,以提高生成中间视图的准确性。分别利用单目图像和中间视图提取单目和双目视觉信息,计算单目质量分数和双目质量分数,并融合得到立体图像的质量分数,达到评价立体图像质量的目的。实验结果表明,MB-FR-SIQA模型在LIVE-I数据库上具有较高的预测精度,其斯皮尔曼等级相关系数、皮尔森线性相关系数、均方根误差分别为0.945、0.951、5.318,且预测的质量分数符合人类主观评估。In the field of Stereoscopic Image Quality Assessment(SIQA),effectively simulate the Human Visual System(HVS) to evaluate the image quality remains an important problem. Considering the visual perception characteristics of human eyes,this paper proposes a SIQA model based on monocular and binocular visual information.The model uses stereo disparity algorithm to obtain disparity matrix of reference and distorted stereo images,and employs the saliency map,Gabor energy and disparity matrix response image to optimize the weight coefficients of the left eye and right eye,and thus improves the accuracy of the generated cyclopean image. Then the monocular and binocular visual information features are extracted from monocular images and cyclopean images respectively. The features are combined with the quality score to evaluate the quality of stereoscopic images. The experimental results show that the Spearman rank correlation coefficient,Pearson linear correlaion coefficient and Root mean square error of the model on LIVE-I reach 0.945,0.951 and 5.318 respectively. The proposed model displays a higher prediction accuracy,and its results are more consistent with human perception.

关 键 词：立体图像质量评价 人类视觉系统 单目视觉 双目视觉 中间视图 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]